Revert #103880 "Use non-ascribed type as field's type in mir"

[lqd]

This PR prepares a revert for #103880 to fix #105809, #105881, #105886 and others (like the duplicates of the first one), in case an actual fix can't get done today.

I've also added the MCVE from #105809. There is no MCVE for the #105881 and #105886 ICEs yet however, so there are no tests for them here, although we'll need one before relanding the original changes.

Were this PR to land, it would also reopen #96514 as it was fixed by the original PR.

Opening as draft to allow time for a possible fix.

r? @jackh726

[jackh726]

@bors p=1 rollup=never

r=me in 6-8 hours if no fix has been put up

[b-naber]

Can you add -Zmir-opt-level=1 here? It was reported in one of the issues that this only ICE'd with that option.

[lqd]

I can, but IIRC 1 is the default, and this test does ICE as is on master.

So I take it we'll land the revert and won't have time to do a fix ?

[b-naber]

I'm looking into it right now, but am not sure whether I can fix it in the next couple of hours. If I don't open a PR in the next two hours we can revert it.

[lqd]

Don't worry about it too much, it's perfectly fine, we can land the revert to buy us time to look into it without pressure.

I've now added -Zmir-opt-level=1 to the test.

[b-naber]

Getting somewhat closer to figuring out what is going wrong. Can you give me 30 more minutes?

[lqd]

Ah but of course! #105881 and #105886 don't have MCVEs yet so hopefully these 3 issues all have the same underlying issue.

I've marked this PR as ready, and if you want it to land, either you or @jackh726 can tell bors that whenever you need. The time frame jack refers to here #105905 (comment) would be 2AM my time.

[b-naber]

We add an additional OpaqueCast here because ty != pattern.ty. This then later ICEs because we cannot get the layout for that type when projecting that in codegen. I don't understand why getting the layout fails though. Will try to work more on this in the next few days.

Thanks for opening the revert PR @lqd.

[rustbot]

This PR changes MIR

cc @oli-obk, @RalfJung, @JakobDegen, @davidtwco, @celinval, @vakaras

[jackh726]

@bors delegate=b-naber

@b-naber From the comment thread, not sure if you're stuck. Feel free to r=me at any point. It's very easy to revert and fix later, so no stress.

[bors]

✌️ @b-naber can now approve this pull request

[b-naber]

@bors r=jackh726

[bors]

📌 Commit 5457db9 has been approved by jackh726

It is now in the queue for this repository.

[lqd]

Bumping the priority to 8, this is a P-critical fix that should land before the #105918 rollup at p=7.

@bors p=8

[bors]

⌛ Testing commit 5457db9 with merge 696563e...

[bors]

☀️ Test successful - checks-actions
Approved by: jackh726
Pushing 696563e to master...

[rust-timer]

Finished benchmarking commit (696563e): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.7%	[-1.3%, -0.3%]	5
Improvements ✅ (secondary)	-1.7%	[-3.5%, -0.4%]	14
All ❌✅ (primary)	-0.7%	[-1.3%, -0.3%]	5

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	3.0%	[3.0%, 3.0%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	3.0%	[3.0%, 3.0%]	1

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.5%	[-2.5%, -2.5%]	1
All ❌✅ (primary)	-	-	0